Landsat IDs tornado damage, and damage recovery in forests is evaluated.
Part one of our study explores how tornadoes affect the reflectance and commonly used indices derived from reflectance, such as the Normalized Difference Vegetation Index (NDVI) and the Tasseled Cap indices of brightness, greenness, and wetness, when comparing damaged forest, grassland, and urban land cover. In part two of this study, we examine five years of Landsat imagery surrounding the 27 April 2011 tornado outbreak to compare a forest disturbance index (DI) to NDVI in and around five tornado tracks.
To explore how tornadoes affect the reflectance and subsequent calculations of NDVI and Tasseled Cap indices, we examine 17 tornadoes in the United States using Landsat 5/7 TM/ETM+ imagery taken within one month after each tornado event. After drawing contours of the damaged area, we used the National Land Cover Dataset to determine the LCLU of each damaged grid cell and sampled an equivalent nondamaged grid cell 10-60 km away. A comparison of the regions reveals that most tornado-damaged surfaces exhibit a higher median reflectance in both the visible and SWIR bands and a lower reflectance in the NIR band. This generally corresponds to a higher Tasseled Cap brightness and a lower, greenness, wetness, and NDVI. This signal was most prominent in areas of homogenous vegetation (i.e., forests) and less prominent within heterogeneous land cover. Tornado strength and season also alter the magnitude of change.
Tornado damaged forests reveal an analogous signature to forest clearing, and many ecological studies use a disturbance index (DI) tuned to this signal to quantify recovery. Prior to the tornado outbreak in April 2011, DI values remain relatively stable, with damaged forests showing an increase in DI values (i.e., more disturbed) after the outbreak. In contrast, NDVI values fluctuate with the seasons, highlighting a difficulty in performing pixel-based comparisons across time. DI is immune to such a natural cycle, making it a promising metric for longitudinal analyses of tornado damage recovery.--Darrel M. Kingfield (University of Oklahoma/CIMMS, NOAA/ NSSL), and K. M. de Beurs, "Landsat identification of tornado damage by land cover and an evaluation of damage recovery in forests," in the April 2017 issue of the Journal of Applied Meteorology and Climatology.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||PAPERS OF NOTE|
|Author:||Kingfield, Darrel M.; de Beurs, K.M.|
|Publication:||Bulletin of the American Meteorological Society|
|Date:||Jun 1, 2017|
|Previous Article:||How to make local context matter in national climate adaptation.|
|Next Article:||On the seasonality of Arctic black carbon.|